Time-varying associations between corticosteroid dose and hospital mortality in ARDS: a sliding-window analysis of MIMIC IV
File(s)
Published version
Author(s)
Marshall, Dominic
Jamme, Matthieu
Patel, Brijesh
Antcliffe, David
Parbhoo, Sonali
Type
Journal Article
Abstract
Background
Corticosteroids are now recommended in guidelines for patients with acute respiratory distress syndrome (ARDS); however, optimal timing and dose remain uncertain. We assessed whether the association between corticosteroids and hospital mortality varies over time in the ICU.
Methods
We performed a retrospective observational study of ARDS patients identified in the MIMIC IV database (2008–2019). To analyze the time-varying association between corticosteroids and hospital mortality, we constructed overlapping three-day windows from ARDS days 0 to 14. We compared windows with no corticosteroid exposure (0 mg) to windows meeting cumulative prednisolone-equivalent dose thresholds chosen to approximate regimens from landmark clinical trials (≥30, ≥150, ≥270, ≥390 mg PE over 3 days). We estimated overlap weighted, doubly robust adjusted risk differences (OWRD) for hospital mortality using augmented inverse probability weighting (AIPW).
Results
Of 987 included patients, 354 (35.9%) received corticosteroids, with 262 (74.1%) and 128 (36.2%) of treated patients meeting the ≥150 mg and ≥390 mg thresholds in at least one window. Early, low cumulative dosing (≥30 mg/3d) was not associated with a detectable difference in hospital mortality (e.g., days 0–2: OWRD 0.03, 95% CI −0.05 to 0.10). Conversely, higher cumulative doses received later in the ICU stay (≥150–390 mg/3d after day 8) were associated with higher observed mortality (e.g., days 8–10: OWRD 0.25, 95% CI 0.11–0.39). However, estimates in late, high-dose windows were less precise due to limited covariate overlap and smaller sample sizes.
Conclusions
In unselected ARDS, we found no evidence of benefit or harm from early lower-dose corticosteroids, but higher cumulative doses later in ICU stay were associated with higher mortality, possibly reflecting residual confounding and limited covariate overlap. These hypothesis generating findings support randomized studies testing corticosteroid timing and dose in ARDS.
Corticosteroids are now recommended in guidelines for patients with acute respiratory distress syndrome (ARDS); however, optimal timing and dose remain uncertain. We assessed whether the association between corticosteroids and hospital mortality varies over time in the ICU.
Methods
We performed a retrospective observational study of ARDS patients identified in the MIMIC IV database (2008–2019). To analyze the time-varying association between corticosteroids and hospital mortality, we constructed overlapping three-day windows from ARDS days 0 to 14. We compared windows with no corticosteroid exposure (0 mg) to windows meeting cumulative prednisolone-equivalent dose thresholds chosen to approximate regimens from landmark clinical trials (≥30, ≥150, ≥270, ≥390 mg PE over 3 days). We estimated overlap weighted, doubly robust adjusted risk differences (OWRD) for hospital mortality using augmented inverse probability weighting (AIPW).
Results
Of 987 included patients, 354 (35.9%) received corticosteroids, with 262 (74.1%) and 128 (36.2%) of treated patients meeting the ≥150 mg and ≥390 mg thresholds in at least one window. Early, low cumulative dosing (≥30 mg/3d) was not associated with a detectable difference in hospital mortality (e.g., days 0–2: OWRD 0.03, 95% CI −0.05 to 0.10). Conversely, higher cumulative doses received later in the ICU stay (≥150–390 mg/3d after day 8) were associated with higher observed mortality (e.g., days 8–10: OWRD 0.25, 95% CI 0.11–0.39). However, estimates in late, high-dose windows were less precise due to limited covariate overlap and smaller sample sizes.
Conclusions
In unselected ARDS, we found no evidence of benefit or harm from early lower-dose corticosteroids, but higher cumulative doses later in ICU stay were associated with higher mortality, possibly reflecting residual confounding and limited covariate overlap. These hypothesis generating findings support randomized studies testing corticosteroid timing and dose in ARDS.
Date Issued
2026-12-01
Date Acceptance
2026-03-23
Citation
BMC Pulmonary Medicine, 2026, 26 (1)
ISSN
1471-2466
Publisher
BMC
Journal / Book Title
BMC Pulmonary Medicine
Volume
26
Issue
1
Copyright Statement
© The Author(s) 2026. This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
License URL
Identifier
10.1186/s12890-026-04251-w
Subjects
ARDS
Corticosteroids
Dose
Timing
Sliding window
Augmented inverse probability weighting
Overlap weights
MIMIC-IV
Observational study
Publication Status
Published
Article Number
ARTN 209
Date Publish Online
2026-03-26